Validation of a computerized decision support system to review pharmacotherapy treatment: scheduling guidelines.

Background: The review of pharmacotherapy can be conceptualized as a service in which the drugs used by the patient are reviewed to control the risks as well as to improve the results of the drug therapy, detecting, solving, and preventing issues associated with the drug, readjusting the doses and times (schedule) so that the treatment is not incompatible or in duplicity.

Methods: The aim of the study was to validate an intelligent information system, which was developed to assist the scheduling activity in the pharmacotherapy review. The system used the concept of Genetic Algorithms. To validate the system, hypothetical cases were elaborated considering various aspects of pharmacotherapy such as underdose, overdose, drug interactions and contraindications. These cases were tested in the system and were also analyzed by pharmaceutical experts with clinical and research experience in the pharmacotherapy review process. The degree of agreement between the assessments of the appointments carried out by the pharmaceutical specialists and by the system were measured using the Kappa index with a 95% confidence interval.

Results: In detecting errors and make propositions, the system was able to identify 80% of errors, with pharmaceutical experts identifying between 20 and 70% of errors. In relation the results of kappa between the cases, the system had 87,3% of concordance, whereas the best pharmaceutical expert had 75,5% of concordance, considering the correct answer.

Conclusion: It can be concluded that with the methodology used, the investigation met the objectives and confirmed the system is effective for pharmaceutical review process. There are indications that the system can help in the Pharmacotherapy review process, being able to find prescription errors as well as to establish times for the use of medications according to the patient's routine.